WhiteFiber, Inc. Ordinary Shares Q1 FY2026 earnings call

Key Operational Milestones

• Completed the company's IPO and transitioned to standalone public company operation in 2025, establishing required independent reporting and governance frameworks.
• Brought Montreal 3 online in Q4 2025: retrofitted an existing industrial factory into a custom data center in 6 months, faster than traditional ground-up development, to serve AI infrastructure firm Cerebrus. The company is exercising its purchase option for the site for ~24 million Canadian dollars, funded via a Royal Bank of Canada facility, which will reduce annual lease payments by ~3.1 million Canadian dollars.
• Signed a 10-year, 40 megawatt IT load contract with Nscale for the NC1 project, representing ~$865 million in total contracted revenue. A customer-requested design modification shifted the ready-for-service date to May 31, 2026, with all associated change costs covered by the customer. Nscale has since secured an offtake agreement with an investment-grade hyperscale end customer, strengthening the project's credit profile.
• Repositioned the cloud segment to focus on enterprise deployments and managed infrastructure, rather than commodity bare metal leasing: monetized ~1,000 H200 GPUs for ~$26 million near cost to redeploy capital, recycled terminated customer capacity into new longer-term contracts, and now holds ~80% of monthly recurring cloud revenue under contract with a 22-month weighted average remaining term. The pro forma GPU fleet totals ~3,700 GPUs across multiple NVIDIA architectures.

Core Strategic Priorities

• Successfully bring NC1 online, secure long-term project financing, and market additional capacity beyond the initial 40 megawatts starting in mid-2026.
• Advance the next site in the development pipeline, targeting at least one additional customer deployment in 2026, with disciplined prioritization of high-quality hyperscale and enterprise counterparties.
• Maintain capital discipline: prioritize opportunities with strong credit profiles, durable revenue, and attractive returns over rapid unprofitable growth. Leverage the company's core competitive advantage of fast execution via retrofit development, which reduces development risk and accelerates time to revenue.
• Evaluate capital-optimizing alternatives for the smaller Montreal 2 asset, including a potential sale or redeployment of capital to larger-scale opportunities.

WhiteFiber reported total fourth quarter 2025 revenue of $23.6 million, up from $20.2 million in Q3 2025 and $14.6 million in Q4 2024. The firm has two operating segments:

1. Collocation (Data Center) Segment: Revenue was $3.9 million, up from $1.7 million in the prior quarter. This increase reflects the partial-quarter revenue contribution from the newly brought-online Montreal 3 facility serving Cerebrus. This segment accounted for approximately 16.5% of total Q4 2025 revenue.
2. Cloud (GPU Cloud Services) Segment: Revenue was $19.3 million, up from $18.0 million in the prior quarter. This segment accounted for approximately 81.8% of total Q4 2025 revenue.

Gross margin (excluding depreciation) improved to 61% in Q4 2025, up from approximately 52% in Q4 2024. Adjusted EBITDA for the quarter was $5.8 million (25% margin), with full-year 2025 adjusted EBITDA of $17.3 million. Operating loss for Q4 2025 was $5.4 million, and net loss was $1.5 million.

• Cloud segment revenue is expected to decline in H1 2026, with Q1 2026 revenue projected between $16 million and $17 million, and April 2026 representing the revenue low point. Revenue is expected to begin ramping in mid-Q2 2026 and accelerate through the second half of the year, with full-year 2026 H2 cloud revenue expected to be substantially higher than H1. Cloud margins are expected to remain consistent during the transition as fixed costs scale down with near-term revenue.
• General and administrative expense for Q1 2026 is expected to be slightly higher than the prior quarter, driven by increased headcount and platform expansion investments.
• NC1 debt financing is now expected to close in Q2 2026, delayed from the original expected timeline due to more rigorous lender underwriting, but the strengthened project credit profile is expected to result in improved financing terms compared to earlier expectations. The company has sufficient existing liquidity to fully fund NC1 completion before permanent financing is secured.
• The company targets bringing at least one new site and customer deployment online during 2026. Additional NC1 capacity marketing is expected to begin around mid-2026, with greater visibility on timing available by that point.

• Hyperscale customer due diligence has extended site selection timelines, as customers have very specific technical and operational requirements that require extensive verification before deals can be finalized.
• Power delivery and substation upgrade timelines for additional NC1 capacity are dependent on Duke Energy's scheduling and equipment availability, which can cause delays beyond the company's control.
• Lender underwriting for data center project financing has become more rigorous and time-consuming, extending financing timelines relative to initial expectations, though the company has bridge financing options to cover near-term liquidity needs.
• Retrofit site pricing has increased modestly as property owners have adjusted prices in response to the AI infrastructure boom, though site availability remains sufficient for the company's current pipeline.
• Enterprise cloud deployments have longer sales cycles, creating a near-term timing mismatch that will result in lower H1 2026 cloud revenue as the segment repositions.
• Greenfield data center development carries higher execution and development risk compared to the company's preferred retrofit model.

Q: What dominoes need to fall to add additional power at NC1, and what is the customer profile for smaller 5-20 megawatt versus large 100+ megawatt deployments?

A: Additional power at NC1 requires Duke Energy to complete scheduled substation upgrades on the site's property, as Duke has already committed to provide the extra capacity. For tranche size, the customer base is largely similar: even large hyperscalers often seek smaller quick-turn deployments in specialized urban locations, and the company's fast execution reputation has increased demand for these smaller sites in the current capacity-constrained market. All customer selections prioritize minimizing counterparty risk and supporting cost-effective project financing to maximize equity returns.

Q: What caused the NC1 change order that delayed the ready-for-service date, and can you share the expected term of Nscale's hyperscale offtake agreement?

A: The change order was driven by the end customer's request for additional networking layout optionality, and has no connection to NVIDIA GPU architecture updates. All incremental capital costs from the change order are passed directly to the customer per the contract terms. The company declined to disclose the term of the offtake agreement, noting it cannot share commercially sensitive details and market participants can reference public market precedent for typical terms.

Q: When the initial cloud customer terminated its contract, how quickly was the capacity replaced, and will the next new 2026 facility be similar in scale to Montreal 3 or NC1?

A: All terminated capacity from the initial cloud customer has already been recontracted: the H100 capacity was placed under a new two-year contract worth $50 million starting mid-April, and the B200 capacity was split across two new contracts with longer durations than the remaining term of the original terminated agreement. The next new facility is expected to be similar in scale to the NC1 project, which uses the company's core retrofit model that reduces execution risk and delivers operational capacity 40% cheaper than ground-up development, a core competitive advantage for the firm.

Q: What does the company's new more deliberate approach to customer announcements entail, following the NC1 process?

A: The new approach waits to make customer announcements until all key commercial and financing components of a project are fully aligned, rather than announcing deals early when multiple workstreams are still progressing at different speeds. The company also will be more reserved about hinting at future contract wins to avoid ceding negotiation leverage with customers and stakeholders. New customer opportunities and site development discussions are run in parallel, with priority given to hyperscale opportunities.